
\begin{table}[h]
\caption{Treatment Effects: Alternative Knowledge Measures}
\begin{center}
\begin{threeparttable}
\begin{tabular}{l@{} D{.}{.}{3.5}@{} D{.}{.}{3.4}@{} D{.}{.}{3.5}@{} D{.}{.}{3.4}@{} D{.}{.}{3.5}@{} D{.}{.}{3.3}@{} D{.}{.}{3.3}@{} D{.}{.}{3.3}@{}}
\toprule
 & \multicolumn{2}{c}{Top 4 LGU/own} & \multicolumn{2}{c}{Top 3 LGU/own} & \multicolumn{2}{c}{Top 4 LGU/their} & \multicolumn{2}{c}{Top 4 CIV/own} \\
\cmidrule(lr){2-3} \cmidrule(lr){4-5} \cmidrule(lr){6-7} \cmidrule(lr){8-9}
 & \multicolumn{1}{c}{(1)} & \multicolumn{1}{c}{(2)} & \multicolumn{1}{c}{(3)} & \multicolumn{1}{c}{(4)} & \multicolumn{1}{c}{(5)} & \multicolumn{1}{c}{(6)} & \multicolumn{1}{c}{(7)} & \multicolumn{1}{c}{(8)} \\
\midrule
Treated      & 0.21^{*}                  & 0.25^{**}                 & 0.16^{**}                 & 0.19^{**}                 & 0.15                      & 0.18^{*}                  & 0.14                      & 0.16                      \\
             & (0.10)                    & (0.10)                    & (0.07)                    & (0.07)                    & (0.10)                    & (0.10)                    & (0.10)                    & (0.10)                    \\
Base Outcome & 0.27^{***}                &                           & 0.21^{***}                &                           & 0.25^{***}                &                           &                           &                           \\
             & (0.03)                    &                           & (0.05)                    &                           & (0.04)                    &                           &                           &                           \\
\midrule
Model        & \multicolumn{1}{c}{Panel} & \multicolumn{1}{c}{Cross} & \multicolumn{1}{c}{Panel} & \multicolumn{1}{c}{Cross} & \multicolumn{1}{c}{Panel} & \multicolumn{1}{c}{Cross} & \multicolumn{1}{c}{Panel} & \multicolumn{1}{c}{Cross} \\
R$^2$        & 0.13                      & 0.07                      & 0.12                      & 0.09                      & 0.11                      & 0.07                      & 0.08                      & 0.06                      \\
Observations & 627                       & 632                       & 627                       & 632                       & 627                       & 632                       & 627                       & 632                       \\
Clusters     & 18                        & 18                        & 18                        & 18                        & 18                        & 18                        & 18                        & 18                        \\
\bottomrule
\end{tabular}
\begin{tablenotes}[flushleft]
\scriptsize{\item \sym{*} \(p<.10\), \sym{**} \(p<.05\), \sym{***} \(p<.01\)} 
\item Panel models include a dummy variable for missing baseline. 
\item All models include station (municipality) and rank fixed effects for the blocking variables 
\item Models labeled with "Top 4" indicate that the outcome was measured as the match of officer responses with the top 4 issues mentioned by LGU/civilian respondents in the same municipality, while models labeled "Top 3" indicate that we matched with only the top 3 issues mentioned. Models labeled "LGU" indicate that we matched officer responses with the aggregate responses of the LGU community leaders, while models labeled "CIV" indicate that we matched with the aggregate responses of ordinary civilians. Models labeled "own" indicate that we matched using the issues that officers themselves thought were important, while models labeled "their" indicate that we matched using the issues that officers thought that LGU/civilian respondents thought were important.}
\end{tablenotes}
\end{threeparttable}
\label{tab:KnowledgeAlt}
\end{center}
\end{table}
